Chandra Devisetti

Engineered distributed ETL pipelines on Databricks for multi-terabyte healthcare datasets, reducing batch processing time by 45%. Designed GraphRAG retrieval using Neo4j knowledge graphs with vector embeddings to improve relationship-aware clinical query resolution by 35%. Built knowledge graph pipelines extracting entities/relationships from unstructured clinical documents, and developed ontology extraction to standardize domain schemas. Implemented production monitoring and failure recovery for data reliability (99.9%) and faster incident resolution. Built high-throughput document ingestion pipelines converting medical documents into vector embeddings for RAG in Pinecone. Orchestrated end-to-end MLOps with SageMaker Pipelines and automated LLM endpoint deployment via Amazon Bedrock with guardrails for safe, grounded answers. Developed backend services using Java/Spring Boot, semantic search with sub-300ms latency, and deployed ML inference to EKS with autoscaling. Integrated HIPAA/HI

Designed AWS Glue ETL to transform/load S3 data into Redshift, improving BI query performance by 35%. Built a financial knowledge graph for fraud workflows and implemented ontology-driven payment data modeling for standardized schemas. Developed REST APIs and microservices using FastAPI, and built real-time ingestion using Kafka with AWS Lambda/Azure Functions to process payment signals in under 200ms. Migrated cron jobs to containerized Apache Airflow on EKS/AKS, reducing failure rates by 50%. Provisioned multi-cloud environments using Terraform for RDS/Azure SQL/S3 and integrated Great Expectations with dbt for validation to prevent dirty data and increase reporting confidence by 70%. Implemented secure access using AWS IAM + Azure AD with Snowflake RBAC to enforce PCI-DSS and GDPR compliance. Built a Data API with FastAPI and Redis improving dashboard load time by 80%. Implemented event-driven reconciliation using S3/Azure Blob events with SQS/Lambda. Automated CI/CD for dbt and Air

Prototyped autonomous AI agents for creative auditing by bridging Azure OpenAI (GPT-4) with Python-driven RPA (Playwright), reducing dependence on human moderators by 50%. Architected a reconciliation engine connecting Node.js services on Azure App Service with Python validators to reduce billing discrepancies by 35%. Migrated legacy cron jobs to containerized Dagster/Docker on EKS/GKE for high-availability reporting and reduced pipeline failure rates by half. Built an API gateway on GCP (FastAPI + Redis) to aggregate MarTech signals and automate bid adjustments 15x faster than manual work. Integrated Salesforce Media Cloud with Google Ad Manager using a Python Order Creator to accelerate revenue transitions by 48 hours. Optimized retail media operations with AWS Lambda + Amazon Advertising APIs, saving 40% effort. Built NLP moderation pipelines using HuggingFace on GCP Cloud Run to improve moderation throughput by 4x. Implemented serverless ingestion using AWS SQS and GCP Pub/Sub for